您好,欢迎访问三七文档
当前位置:首页 > 商业/管理/HR > 企业财务 > 用ROC曲线确定最佳临界点和可疑值范围
【】[],,610041[](1976-),,,.用ROC曲线确定最佳临界点和可疑值范围陈卫中,倪宗瓒,潘晓平,刘元元,夏彦[] :ROC,,,ROC。:508176,ROC。:ROC0.9723,,。ROC,40.70U/L,,(36.90,47.80)U/L,。:ROC,。[] ROC;;;[]R541.4 []A []1003-8507(2005)07-0729-03RECEIVEROPERATINGCHARACTERISTICCURVESTODETERMINETHEOPTIMALOPERATINGPOINTANDDOUBTABLEVALUEINTERVAL.CHENWei-zhong,NIZong-zan,PANXiao-ping,etal.WestChinaMedicalCenter,SichuanU-niversity,Chengdu,610041.Abstract:Objective:ToevaluatetheclinicalvalueanddiscriminatorypowerofadiagnostickittodiagnoseCoronaryHeartDiseaseanddeterminetheOptimalOperatingPoint(OOP)usingtheReceiverOperatingCharacteristic(ROC)curves,andexplorethemethodtodeter-minethedoubtablevalueinterval.Methods:BasedontheCAGandclinicalobservations,508peoplewiththediseaseand176withoutthediseasewereselectedtomeasuretheirOxLDL,andtheROCcurvewasplotted.Results:Theareaunderthecurvewas0.9723withahighstatisticalsignificance.ItshowedthatthediagnostickitwassuitableforCoronaryHeartDiseasescreening,theOOPofthediagnostickitwasdeterminedandinterpretedat40.70U/Lwithhighersensitivityandspecificity.Thedoubtablevalueintervalof36.90U/Lto47.80U/Lwasdeterminedatthesametime.Conclusion:ItwasusefulusingROCcurvewhenassessingtheperformanceanddeterminingtheOOPanddoubtablevalueintervalofascreeningtest.KeyWords:OOP;ROC;Doubtablevalueinterval;CoronaryHeartDisease ROC(receiveroperatingcharacteristic),,1-,,、[1]。ROC[2]。ROC,,,。,ROC,,Youden[3],,,。,,,,,。,。,,。1 1.1 ,707,508,18~70;176,143,33;()23。1.2 ROC,Youden,2[4],23,、。2 2.1 。,。1。,。1 (U/L)50820.1815.0215.55131.905.202344.1311.9043.5079.4020.8017668.9018.4268.05120.8015.8070733.0926.3521.40131.905.202.2 ROC ,,,(1-)ROC(2)。,7292005327 ModernPreventiveMedicine,2005,Vol.32,No.7。0.9723,,Az=0.5,,。2.3 ,。,,Youden,,≥40.70U/LROC(2A)。96.02%,93.50%。1 ROC2 ROC95%LCL95%UCLZP0.97230.00620.96020.984476.55090.00002.4 ,X,,,0.88,1-0.04,,,,0.98,1-0.098,。0.880.98,。(2B、C)36.9047.80U/L。2.5 3 (36.90,47.80)U/L,3,χ2=585.4066,P0.0001,,(P)=0.6730。Kappa=0.7576,。3。3 3n36.9036.90~47.8047.8050846391.14244.72214.13176521.741252.17626.092331.701910.8015487.5070747166.62557.7818125.603 ROC[5~9]。ROC(Az)。0.51。Az0.5~0.7,;0.7~0.9,;0.9。ROCAz=0.5,,,。,0.9723,,Az=0.5,,。,ROC,,Youden,40.70U/LROC。(96.02%)(93.50%),。ROC,,,Youden,,,,,。,,、,,1。(),[6],,。,;,。,。,,,。,,100%。ROCROC,。,。,,[3]。,,,,,,,X,,100%,,,2[4],23,、。,36.9047.80U/L,,Kappa0.7576,0.75,,。,,ROC,ROC,,7302005327 ModernPreventiveMedicine,2005,Vol.32,No.7,。[][1]GregoryMark.Receiver-operatingcharacteristic(ROC)plots:Fun-damentalEvaluationToolinClinicalMedicine[J].ClinChem,1993,30(4):561-567.[2],,.(ROC)[J].,2001,15(3):201-202.[3],,.ROC[J].,1999,24(9):993.[4].SPSSROC[J].,2003,19(9):1152.[5],.ROC[J].,1999,17(2):66.[6].[M].:,2003.[7]HanleyJA,McNeilBJ.Themeaninganduseoftheareaunderare-ceiveroperatingcharacteristic(ROC)curve[J].Radiology,1982,143(4):29.[8]SwetsJA.Measuringtheaccuracyofdianosticsystem[J].Science,1988,240(4622):1285.[9]HanleyJA,McNeilBJ.Themeaninganduseoftheareaunderareceiveroperatingcharacteristic(ROC)curve[J].Radiology,1982:29-36.[]2004—10—19【】[],,455000[](1966~),,,.公共场所从业人员SARS相关知识调查韩文霞,赵亮[] ;;SARS;[]R563.1 []B []1003-8507(2005)07-0731-011 1.1 ,384,158,41.1%,226,58.9%;173,45.1%;、136,35.4%;43,11.2%;32,8.3%。1.2 ,“SARS”,,35。,,,,,389,384,98.7%。2 2.1 SARS SARS,,100%,98.1%。SARS2~14d73.7%,SARS1~2100%,SARS93.6%,SARS78.6%。SARS100%,SARS100%。17.2%SARS,40.4%,SARS,、45.6%。2.2 SARS 20.6%SARS,;16.9%SARS。100%,34.6%;89.6%,,、。96.9%。SARS(%);SARS,100%。SARS80.5%。SARS12100%,SARS100%,,SARS96.6%。2.3 SARS 100%SARS,。SARS,100%SARS。84.6%SARS。70.6%SARS,6h,12h。2.4 SARS SARS:、、、、54.7%、47.1%、28.4%、32.0%、26.8%29.4%。2.5 SARS 100%,SARS97.7%,SARS86.9%,,78.1%,87.3%,100%;,“”100%,SARS96.2%,,SARS78.6%;SARS,92.2%;78.1%,34.6%、,38.6%。、,86.1%,13.9%。(771)7312005327 ModernPreventiveMedicine,2005,Vol.32,No.7
本文标题:用ROC曲线确定最佳临界点和可疑值范围
链接地址:https://www.777doc.com/doc-7297146 .html